Storage devices (e.g., disk arrays) may be accessible by a server over a network, such as a storage area network (SAN). The server's connection to the storage device may rely on the availability of a variety of components. By way of example, these components may include interfaces of the server (typically interface cards), gateways to the network, and switches.
Failure of any of the components that are used to create the connection between the server and the storage device(s) may result in loss of access to one or more storage devices on the network. Data transfers that are scheduled to take place (e.g., backups) may be disrupted. The device(s) may remain unavailable until a new path to the device is manually established or the failed component is replaced.
Even while the components remain operational, the network on which the storage device(s) reside may exhibit poor performance during heavy traffic periods. The network connections from server(s) to storage device(s) may have one or more bottlenecks in which all communications are routed through a single component, such as a switch. The component at the bottleneck point must be responsible for the entire communication load between the server(s) and storage device(s).